[vtkusers] vtkCellLocator

[John Biddiscombe]

CellLocator should work fine. (The tolerance is used, but the details are 
hidden away deep inside the vtkCell subclasses (polygon, triangle etc etc) 
where it controls line intersection tests - although you didn't specify how 
you are using the locator).

If you use an implicit function to describe a polyhedral radar beam (or 
even a cone) and intersect that with the Dataset using a celllocator you 
can do what you want. [You will get the front and back of a hill though. If 
you are implementing a full backscatter type problem then you can take the 
clipped dataset, run it through a normal generator, and using the antenna 
as a reference then reject the rear facing polygons and compute field 
strength at the remaining points. use these asscalars on the original 
dataset and you get some nice piccies]

JB


At 13:41 10/09/2001, James A Galbraith/GALBJA/CC01/INEEL/US wrote:
>Good morning everyone.
>
>We are modelling radar beams over terrain using vtk with java.  In order to
>mask the radar beams where they intersect the terrain we need to determine
>where this intersection occurs.  We are currently using the vtkCellLocator
>class with mixed results.  For the most part, we are receiving what we
>consider valid intersections, but occasionally we receive intersections
>that are not correct.  In some instances the intersection comes at the back
>side of a hill the radar beam is "passing through" rather than the front
>edge.  In other instances, the beam is allowed to pass through several
>hills, intersecting on , say, the third hill.  Instances where the beam is
>allowed to pass through several intersections before arriving at the "best"
>intersection are very consistent and are always returned this way.  When we
>overlay radar beams on the terrain, the terrain then masks the radar beam
>wherever it is above the beam - which supports our view that we are getting
>invalid intersections from vtkCellLocator.
>
>Can anyone shed some light on how the vtkCellLocator works and why it
>possibly is obtaining a "best" cell which is incorrect as far as we are
>concerned?  Are we using the correct class for detecting these locations,
>or is there another class we should look at?
>
>I tried tweeking the tolerance value, but after looking at the source it
>doesn't appear to be used.
>
>Thanks again for all your help
>
>James Galbraith
>Idaho national Engineering & Environmental Laboratory (INEEL)
>Bechtel BWXT Idaho, LLC
>P.O. Box 1625
>Idaho Falls, ID  83415-3779
>Email: galbja at inel.gov
>
>_______________________________________________
>This is the private VTK discussion list.
>Please keep messages on-topic. Check the FAQ at: 
><http://public.kitware.com/cgi-bin/vtkfaq>
>Follow this link to subscribe/unsubscribe:
>http://public.kitware.com/mailman/listinfo/vtkusers